Ontologies for spatial reasoning, action and interaction

bouncerarcheryΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 1 μήνα)

46 εμφανίσεις

Ontologies for spatial reasoning,
action and interaction



Basic problem statement, techniques under
development, and plans

John Bateman & Till Mossakowski

University of Bremen

NIST Discussion. Tuesday 14th March 2006

http://www.sfbtr8.uni
-
bremen.de

http://www.fb10.uni
-
bremen.de/ontology

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

2

Problem focus


Spatial assistance systems


Route planning and navigation


Real
-
world environments

involving ‘common
-
sense’ entities


Interfacing with geographic information


Interfacing with language technology


Interfacing with visual presentations (maps)


Interfacing with robotic sensor data


Embodied systems


Human
-
Robot Interaction

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

3

Qualitative Information

“In front to the right is
the seminar room”

Quantitative information

Symbolic information

[door_1 recognized]


Bremen Autonomous Wheelchair

:
Rolland

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

4

Basis for the use of ontologies
and ontological engineering


High degree of interoperability between
diverse knowledge
-
rich systems



knowledge of the human world

(commonsense)


knowledge of the robot world

(programmed, emergent)


geo
-
knowledge
(GML, other standards)


spatial knowledge
(spatial calculi)


knowledge of language
(Generalized Upper Model)

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

5

Fundamental issue


The ontologies present are diverse:



different methodologies


different motivations


different domains of application


different worlds


different purposes


different communities

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

6

Methodological starting point


There is no sense in which a simple
‘merging’ of the ontologies involved is a
sensible strategy to follow

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

7

Many perspectives on ‘reality’:

many ontologies

event

time

space
-
1

space
-
2

event

Ontologically diverse

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

8

Ontological diversity


inter
-
ontology mappings

Way description

time

landmarks

choremes

event types

CASL

CASL

CASL

route graphs

CASL

CASL

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

9

Essential properties

we are currently developing


Perspectivalism


Objects


Activities


Artifacts:
spatial artifacts


Language



Granular partitions



Plug
-
and
-
play spatial theories



I1
-
[OntoSpace]

Bateman/Mossakowski 2006

10

Essential ingredients

we are drawing on


Existing ontologies



DOLCE

(for cross
-
category binding and axiomatization)



BFO

(for sites, niches and places and for SNAP/SPAN)



GUM

(generalized upper model for linguistic semantics)

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

11

Essential ingredients

we are drawing on


Formal and computational tools



CASL

Common Algebraic Specification Language

(for specification, structuring and relating)



HETS

Heterogeneous Tool Set

(for connecting to a range of reasoners)



sublanguages of CASL

(e.g., CASL
-
DL, modal CASL)



OWL
-
DL


I1
-
[OntoSpace]

Bateman/Mossakowski 2006

12

Formalization choice: CASL

Common Algebraic Specification Language


de facto
standard
for specification of functional
requirements in software development


developed by the “Common Framework Initiative”
(COFI), an
open

international collaboration


approved by
IFIP WG 1.3

“Foundations of Systems
Specifications”


extensive
User Manual

and
Reference Manual

now
available from Springer (LNCS 2900, LNCS 2960)


I1
-
[OntoSpace]

Bateman/Mossakowski 2006

13

CASL language constructs


Basic specification:

spec

SpecName = Spec


Extension:


Spec1
then

Spec2


Union:



Spec1
and

Spec2


Translations:


Spec
with

SymbolMappings


Parameterization:





spec

Spec1 [Spec2] = Spec


Views:

view
View
:
Spec1
to
Spec2
=
SymbolMapping


(
theory morphisms
)



I1
-
[OntoSpace]

Bateman/Mossakowski 2006

14

Example: PSL specification ...

spec

PSL_subactivity =



PartialOrder with __<=__


s畢慣瑩vi瑹Ⱐ䕬敭e


慣瑩vi瑹

then




... %% axioms for discreteness

end



Michael Gruninger
(
http://www.mel.nist.gov/psl/psl
-
ontology/part12/subactivity.th.html
)


subActivity
: This relation is isomorphic to a discrete partial ordering on the
set

of activities.


PSL_subactivity

PartialOrder

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

15

CASL sublanguages and
environment

Static Analysis

+

Tools

Static Analysis

+

Tools

basic specs

architectural

Theorem Provers


Isabelle

SPASS

...

signatures

development graphs

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

16

Lüttich & Mossakowski

(FOIS 2004)

Axiomatized Ontology in CASL

GenParthood

Primitives

DOLCE

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

17

Lüttich & Mossakowski

(FOIS 2004)

GenMereology

GenParthood

DOLCE

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

18

spec

MEREOLOGY =


PRIMITIVES

then

%%
Ad7, Ad8, Ad9 and Ad10 are generated by

%%

instantiation of GenMereology


GENMEREOLOGY [
sort
T
]

then


GENMEREOLOGY [
sort
S
]

then


GENMEREOLOGY [
sort
PD
]

end

Lüttich & Mossakowski

(FOIS 2004)

GenMereology

GenParthood

Primitives

Mereology

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

19

Lüttich & Mossakowski

(FOIS 2004)

Development Graph


showing dependencies
between specifications

and proof obligations

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

20

The DOLCE ontology in CASL

spec PreDolce =



Mereology_and_TemporalPart


and Temporary_Mereology


and Participation


and Constitution


and Dependence


and Direct_Quality


and Temporary_Quale


and Immediate_Quale

end


spec Dolce =


PreDolce

and


Taxonomy

end

work continuing...

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

21

Ontology construction


Axioms are grouped into logically appropriate theories


Theories may be extended via parameterization to
achieve semantic re
-
use


Theories may be created and related by views: theory
morphisms


Only with this availability of working with
meaningful interrelationships can the complexity
of distinct axiomatized ontologies really be
harnessed.

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

22

Generalized Upper Model : Version 3
(2004
-
)

220 classes

86 properties

OWL
-
DL (
ALCHN
)

disjoint categories iff there
is a specifiable difference
in linguistic reflexes
(grammaticized semantics)

Methodology

Penman Upper Model (1989)

Merged Upper Model (1994)

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

23

The Generalized Upper Model


continues to be used for interacting with our natural language
components. Because of the link to language, it is relatively
straightforward to understand (continuing development since 1985).


Until the beginning of the current project in 2002, it was under seriously
axiomatized.


We are now in the middle of a complete update with axiomatization
and explicit links to DOLCE (via D&S and quality spaces)


note that this does
not

mean that it becomes
merged

with DOLCE!


Work for next 4 years: completion of the axiomatization in the spatial
area, relation to FrameNet and EuroWordNet. Perhaps to WordNet (via
OntoWordNet and SUMO).


Relation to proposals for
simple

Common Subset?

I1
-
[OntoSpace]

Bateman/Mossakowski 2006

24

Summary of work in progress:
with interest in cooperative development


comprehensive formalization of spatial calculi


correctness of composition tables


theory morphisms among different calculi


inheritance of tools along theory/logic morphisms



formal integration of ontologies


via colimits of theories


consistency of integrated ontologies



content development and interrelation of
ontologies


I1
-
[OntoSpace]

Bateman/Mossakowski 2006

25

Approaches to

‘simplifying’ the ontologist’s life...


Making sure that each component of a library of theories only
specifies the axioms which are relevant at that point

(cf. John Sowa: “That is the whole point of Ockham's razor:


eliminate any axioms that are not
absolutely essential to the task at hand.”)



Making sure that unnecessary detail is hidden in ‘upstream’
libraries: CASL



Possibilities for ‘common subsets’:


packages such as our spatial calculi


packages such as DOLCE’s ‘constitution’, ‘participation’, ‘quality
spaces’, BFO’s ‘sites’


language
-
based generic ontology (GUM)